Course Description

Artificial Intelligence (AI) has potential to transform society. In this course you will learn the statistical and computational foundations of different forms of AI, with a particular focus on generative AI that powers ChatGPT and tools that automatically generate images and videos. We will explore the ongoing and potential future impact of this technology on law, science, engineering, business, politics, medicine and education from guest speakers in these domains. We will explore the ethical implications of this technology and approaches to use AI to create a more just society. We will explore questions such as: how can AI be used by doctors to improve medical diagnosis and communication with patients without perpetuating bias that pervades the medical data that AI is trained on? How can AI tutors improve educational opportunities for everyone without depriving students from learning fundamental knowledge that is now so easily copied from large language models? How can AI help us to engineer safer and more environmentally friendly infrastructure for transportation and energy when AI itself is such a large consumer of energy and water?

Texts

We will read excerpts from several books including

Unmasking AI by Joy Buolamwini (2023)

Artificial Intelligence, A Guide for Thinking Humans by Melanie Mitchell (2019)

The Age of AI: And Our Human Future by Henry Kissinger, Eric Schmidt and Daniel Huttenlocher

We will read journal articles, for example,

On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?  By Emily Bender, Timnit Gebru, Angelina McMillan-Major and Shmargaret Shmitchell (2021)

Because AI is changing so rapidly, we will also read contemporary blog posts, popular press pieces and and opinion articles from experts in this field, for example, https://erictopol.substack.com/p/a-big-week-in-medical-ai summarizes major advancements in medical AI each week such as: AI for real-time patient-AI healthcare conversations, AI discovery of new antibiotics, new approaches to mitigate AI bias and analysis of how AI can propogate medical misinformation.

Requirements

Students will attend a weekly 3 hour presentation and discussion seminar. Approximately half of the seminars will be led by guest speakers who are experts at the intersection AI and other fields of science, engineering, medicine, law and ethics. Students will be expected to have read assigned materials before class, actively listen to presentations, engage with speakers with questions, and actively participate in class discussions.

Students will follow a set of step-by-step tutorials to learn how to implement and train AI tools. No prior programming experience is expected or needed; however, students with programming experience can expand upon this foundation for their research papers.

Students should expect most of their out of class time to be dedicated to writing an original research paper on AI and Society. Students will present an oral and written project proposal, mid-semester status update and a final report to the class. Students will be encouraged to use generative AI tools to assist them on their reports and presentations.

About the Instructor(s): Melanie Moses

Melanie Moses began her academic career in an interdisciplinary major called Symbolic Systems at Stanford University where she studied artificial intelligence, robotics, and building computational agents. She has worked in computer translation of natural languages, computer security, modeling of complex systems like the immune system and bio-inspired compution including building “swarms” of robots that forage like ant colonies. She recently co-founded the Interdisciplinary Working Group on Algorithmic Justice with faculty from UNM and the Santa Fe Institute to help policy makers in New Mexico understand how to leverage the benefits and mitigate the risks of Artificial Intelligence.